Promoting a Flourishing Society
 

The Lasting Impact of High School on Giving and Volunteering in the U.S.

Jonathan Schwarz and David Sikkink investigate if high schools in the United States foster behavior, attitudes, and identities that support volunteering and giving among their graduates.

February 1, 2017

Do high schools in the United States foster behaviors, attitudes, and identities that support volunteering and giving among their graduates? Some research has found both short and long term effects of high school experiences on volunteering and community service (Grimm et al. 2005, McFarland and Thomas 2006). Research on civic education and civic participation has generally shown positive effects for private schools in the U.S. (Campbell 2001a, Galston 2001, Niemi and Junn 1998, Wolf et al. 2001, Wolf 2007).

The existing literature, however, does not tell a consistent story on whether religious schools have a positive impact on civic engagement. In this study, we find evidence of a religious school advantage in civic involvement and of substantial religious school influences on the volunteering and giving priorities of their graduates. Other research has criticized Evangelical Protestant schools for being dogmatic, anti-democratic and separatist, unprepared for civic participation in a democratic society (Apple 2001, Peshkin 1986, Rose 1993). Some findings have been more positive (Godwin, Godwin and Martinez-Ebers 2004, Pennings et al. 2014, Sikkink 2009, Vryhof 2004, Wagner 1990), including an important study arguing that Protestant school graduates are more likely to persist in volunteering during their young adult years (Hill and den Dulk 2013). Research on the longer term influence of Catholic schooling on volunteering show mixed results (Dee 2003b, Dee 2005, Dill 2009, Nghiem et al. 2015, Sander 2001). With few exceptions, we find convincing support for the more optimistic view of religious schools, showing that they produce citizens active in civil society.

In this report, we bring important data to the question of the longer term impact of religious high schools on two aspects of civic engagement, organizational volunteering and giving in adulthood. Besides the question of whether school sector influences commitment to volunteering and community service, we address in this report the types of volunteering and community service taken up by graduates of public and private schools. One of the key questions in this area is whether religious schools facilitate volunteering and giving primarily for religious organizations and causes or whether observed sector differences are spread more broadly across community causes or organizations.

School sectors may be more or less effective in shaping volunteering and donor commitments and identities of students. We bring new evidence for this through a comprehensive, national dataset that includes an extensive module on volunteering and giving in adulthood. Our analysis compares graduates from the relatively large private sectors in the U.S.—the Catholic, non-Catholic religious, and nonreligious private sector—to the public school sector. We expect that schools in each sector have strengths and weaknesses in encouraging their students to volunteer and give (Campbell 2001b, Campbell 2008, Sikkink 2009), and that this impact may have effects on students that carry over into their adult years (Dill 2009).

One goal of this study is to test for sector differences that may be a function of variation in the culture of the school. Aspects of school culture embodied in the organization and practices of the school likely influence moral and civic formation of students. Organizational culture—or culture as lived within the school—shapes whether schools are more or less effective at facilitating volunteering and giving, and generating identities and personal commitments to sacrificial action on behalf of others that extend into adulthood. Schools are increasingly interested in community service as a tool for graduating students who are more civically engaged in their communities after graduation (Kleiner and Chapman 1999). The question remains whether this school practice influences adult civic outcomes. Among public schools, there is mixed evidence of the effectiveness of mandatory service through school on later volunteerism (Helms 2013). In other words, requiring students to serve their communities does not make them more likely to voluntarily do so later in life. Are religious schools any different on this score?

This report is a preliminary step in identifying educational environments that improve civic engagement among high school graduates. We offer theoretical reasons to expect that the type of school matters for civic engagement, and we test for differences by school sector, but the data used for this report is not adequate to evaluate the longer term effect of specific schooling practices on giving and volunteering. If graduates of certain schools volunteer or donate money at a higher rate, further empirical study of the mechanisms that link school experiences in religious schools and civic engagement will be warranted. As with most studies in this area, we rely on theories of school sector differences and consider whether our findings support a particular theory or its alternative. After developing ideas on school sector effects in the next section, we lay out our hypotheses and test them against the data available in the PSID. The final section of the report rehearses the evidence for a religious school advantage as well as public school strengths in fostering giving and volunteering in graduates.

Theories of School Sector Differences

Current research provides insight into why we expect differences in civic outcomes by school type. Public schools should have an advantage in fostering positive engagement in community affairs since civic education has historically played a major role in the mission and purpose of public schools (Butts 1978, Butts 1980, Butts 1989). Since the legitimacy of public schools rests in part on their ability to form citizens who participate effectively in democratic institutions, we expect that public schools would, on the whole, attempt to integrate civic engagement into student learning and experiences in high school. This is borne out in the increased focus on community service in public education over the last 30 years (Kleiner & Chapman 1999). As social network hubs in local communities, public schools could have dense networks of strong organizational ties to nonreligious civic organizations in the community. These linkages may facilitate volunteering for public school students by introducing students to these organizations during high school, which could foster a continuing connection through the life course (Grimm et al. 2005). While theoretically convincing, recent studies (Wolf 2007) problematize this view of civic engagement and public schools.

Public schools may foster a sense of obligation to the community in part because of the structure of public education, which is embedded directly into what is considered the public or civic sphere (Gutmann 1987, Macedo 2000). However, social change may have disrupted the connection of public schools and civic life. The relation between public schools and residential or geographic communities is weakening (Coleman and Hoffer 1987). The historic public school commitment to organizational civic engagement may have enervated as schools face financial constraints, academic demands, and competitive challenges from emerging schools of choice that have forced schools to focus inward to meet demands of academic proficiency or college placement (Butts 1989, Lubienski and Lubienski 2014, Noddings 2015). An emphasis on test scores and college admission may preclude public school efforts to build civic commitments or volunteering identities. Further, some have argued that the organizational culture of public schools no longer has the capacity to shape positively the character or moral commitments of students (Arum 2003, Damon 1995, Etzioni 1994, Grant 1988, Hunter 2000). Other than civic skills and connections, which may be generated through public school curriculum and instruction and perhaps through organizational ties established during high school, the public school alumnus, some would argue, is not likely to be particularly well-equipped to continue volunteering beyond high school (Galston 2007).

Catholic schools have several advantages in encouraging volunteering and community service among students. The well-documented religious mission of service to humanity generates strong cultural supports for volunteering and community service in post-Vatican II Catholic schools (Bryk, Lee and Holland 1993). This is reflected in schooling practices, curriculum, and structural links with organizations beyond the school (Meidl and Meidl 2013, Willems et al. 2010). Many Catholic schools attempt to integrate principles of “Catholic Social Teaching,” which includes curriculum and activities meant to foster students who are concerned about poverty and the role of social structures in inequality. Catholic schools are particularly strong in integrating community service and civic engagement directly into classes (e.g., service learning courses) (McLellan and Youniss 2003, Sikkink 2004). Moreover, Catholic schools likely have organizational networks that may facilitate student volunteering and community service, including ties to Catholic Relief Services, Catholic hospitals, and other Catholic institutions active in the civic sphere. The organizational ties to Catholic civic organizations may also have staying power; well-worn pathways to the community may facilitate higher levels of civic engagement into the young adult years.

In addition to these pedagogical practices and structural linkages with other organizations, there is evidence that Catholic school culture embeds a communal orientation (Bryke, Lee, and Holland 1993). Social capital built in these schools, including dense and multidimensional relationships among students, teachers, and parents, which creates a context of reciprocity and trust. The extent that Catholic schools influence formation of students may be higher given the kinds of social capital built within the school community, including a strong sense of belonging and commitment among teachers and students. Parents chose to send their children to Catholic schools, so the schools themselves become elective communities in which families invest their time—parents volunteer their time at high rates in Catholic schools (see Bryk et al, 1993; Carbonaro 1998). The overlap of family and school social worlds may provide a stronger context for student formation toward prosocial goals. Additionally, teachers are more oriented toward communal practices, focusing more on building a strong sense of community-focus within the school (Bryk, Lee and Holland 1993; Hallinan 2008). Further, we would expect that the culture of Catholic schools inculcates “volunteer identities” (Hitlin 2003, Musick and Wilson 2008, Wilson and Musick 1999) that are rooted in a religious community and expressed within a meaningful religious discourse. In short, for Catholic school graduates, some forms of volunteering become a salient part of their identity, which follows them into adulthood. This has greater potential to influence volunteering behavior over the life course.

Less is known about Evangelical Protestant schools. In many ways, Evangelical Protestant schools are likely to be socially similar to Catholic schools (Vryhof 2004), so we would expect them to have the civic advantage of a strong community and collective identity, and organizational ties to local churches. At the same time, ties to civic organizations—especially secular organizations but also to some extent Catholic or mainline Protestant organizations—are likely not as strong in the Evangelical Protestant school sector (Sikkink 2004, Sikkink 2012). A lack of bridges to the civic sphere outside of congregations may reduce the extent of volunteering opportunities for Evangelical Protestant school students. In addition, the mission of Evangelical Protestant schools often focuses on Evangelical religious and spiritual formation of students (Revell 2008), which may dampen volunteering and community engagement through a crowding out process or as a result of efforts to build an intensive community with clear boundaries with the outside community. In contrast to Catholic schools, which tightly link spiritual formation and volunteering identities, Evangelical Protestant orientations to spiritual formation may not have a natural fit with volunteering outside the religious or school community. These schooling experiences may continue to influence civic orientations into young adulthood, partly because of the lack of bridges to dominant civic organizations.

Despite these potential roadblocks, there are also theoretical reasons to expect higher levels of civic engagement among Evangelical Protestant school graduates. Evangelical Protestant school student volunteering may persist into young adulthood because these schools instill religious commitments to meet human needs through personal sacrifice. And ties to congregations, which may be enhanced through Evangelical Protestant school experiences, may provide organizational pathways to volunteering in the adult years (Hill and den Dulk 2013).

Questions and Hypotheses

The question for this report is whether high school experiences have a lasting effect on civic engagement. After leaving high school, are alumni/ae still influenced by school experiences toward volunteering or giving? The long reach of school influence may be difficult to detect. First, the school effect may be indirect, operating through sector effects on educational attainment or religious involvements, for example. Over time, these school effects may be masked by differences in religious practices or educational credentials. Second, a school that facilitates volunteering during enrollment may enhance civic know-how and ties to civic organizations, but this knowledge-base might erode over time due to changing contexts across the life course, such as college, career, or family life. How would this vary by school sector? Are some kinds of schools more effective in influencing volunteering and giving in adulthood?

Our first set of hypotheses regard Catholic schools. We expect that Catholic schools have stronger organizational and cultural supports that facilitate volunteering for their students than public schools. This schooling environment contributes to civic formation that includes higher levels of giving and volunteering in adulthood. The alternative is that the trend toward a college preparatory focus of many Catholic schools (Baker and Riordan 1998, Youniss and McLellan 1999), along with processes related to the educational and career trajectory of their graduates, may dampen the impact of Catholic school attendance on volunteering and giving in adulthood.

Our second set of hypotheses includes the expectation that Evangelical Protestant schools provide stronger supports for volunteering over the life course than public schools. The persistence of volunteering into adulthood may depend on stronger volunteering identities and commitments that are fostered in Evangelical Protestant schools. Despite a relative lack of organizational bridges to civic life, the Evangelical Protestant school graduates may have come to value sacrificial commitments to meeting human needs, which may be expressed through giving and volunteering. The alternative hypothesis is that Evangelical Protestant school graduates are less active in volunteering and giving since, in comparison, public schools have well-established bridges to civic organizations and at least the remnants of their historic mission of civic education. That gap may be wider given the growing emphasis on service learning in public schools (Galston 2001, Kleiner and Chapman 2000).

Our final set of hypotheses focus on the patterns of donating and volunteering across school sectors, which likely reflect the extent and nature of ties between schools and civic organizations. Evangelical Protestant school ties to religious organizations may continue to funnel their graduates giving and volunteering primarily in and through religious causes and organizations, especially congregations. The experience of commitment and belonging to an Evangelical Protestant school, in which students practice sacrifice for others and teachers model the value of commitment to a religious mission, prepares students to make adult commitments within the Evangelical Protestant community, particularly within Evangelical Protestant congregations. Given the Evangelical Protestant focus on family and youth formation (Smith and Snell 2009, Wilcox 2004), we expect a higher level of civic commitment in the areas of youth, families, and education among Evangelical Protestant school graduates. Compared to Evangelical Protestant schools, we posit that Catholic school organizational ties are more general, though they may be focused on civic sectors in which Catholic organizations are well-represented, such as health care and social services. Post-Vatican II Catholic school graduates likely have greater opportunities and interest in volunteering and giving in areas of health care, education, and poverty relief. The lower boundaries between Catholic and nonreligious civic organizations is likely to lead to a balance between religious and nonreligious organizational involvement for Catholic school graduates.

There is less research explicitly on nonreligious private school graduates and adult civic engagement, though one longitudinal study found that a positive private school effect on volunteering and voter participation in adulthood was primarily driven by nonreligious private school graduates (Dill 2009). We hypothesize that nonreligious private school graduates are more likely to volunteer than public school graduates, and that their civic involvement reflects the concentration of higher SES families in private non-religious schools. As such those who attended private non-religious schools will be more engaged with the arts and cultural organizations. Their experiences at secular schools are not likely to lead to interests in, or connections to, volunteering or giving at or through religious congregations. More secular causes and organizations are likely to receive more attention from nonreligious private school graduates. Similarly, public school graduates would be more likely to volunteer and give through nonreligious organizations, since these bridges may have been built through their high school experiences.

Data and Methods

For this report, we draw on one of the most comprehensive longitudinal studies conducted in the United States, the Panel Study of Income Dynamics (PSID). PSID is the world’s longest active panel study of households and families. In 1968, researchers randomly sampled 5,000 households in the United States, collecting data on each family member. This process yielded a baseline sample of over 18,000 individuals. As members of this sample progress through life, new members of their families—spouses and children—inherit the “PSID gene” and are incorporated into the sample. As of 2013, the PSID sample includes information on over 75,000 individuals who are surveyed every other year. Our working sample includes respondents aged 18-88 years.

To test for sector effects in volunteerism and philanthropy, we incorporate information about individuals collected across multiple waves. For instance, the 1995 household survey provides the most detailed information about primary or secondary school sector, the 2011 household survey features the most contemporary information on religious preference, practices, and volunteering, while the 2013 household survey contains the most up-to-date information on philanthropy. For demographics (age, race, region, and urbanicity), we use the most recent response available. For many respondents, demographic data was collected in 2013, though for a few we had to look all the way back to the 1985 household survey for the most recent information.

The PSID survey asked respondents whether they volunteered for a variety of civic organizations during the 2012 calendar year. For example, the questions include whether the respondent did volunteer work in or through a religious congregation:

During [YEAR], did you yourself do volunteer activity at or through your church, synagogue, or mosque, such as serving on a committee, assisting in worship, teaching, or helping others through programs organized by your place of worship? Please do not include volunteering through schools, hospitals, and other charities run by religious organizations. I will be asking you about that volunteering next.

Subsequent questions ask the respondent about other categories of volunteering, reporting on volunteering beyond what had been covered earlier in the survey (“Not counting your volunteering you just told me about…”). Categories include organizations that help children and youth, senior citizens, people in poor health, and people in need. The final categories are organizations that work for social change, which we do not include in our discussion here, and “other volunteering.” Note that this line of questioning explicitly focuses on formal volunteering through organizations, and asks respondents not to include informal volunteering and helping activities.

For the purposes of this report, we consider only the first category, volunteer work at or through a religious congregation, as volunteer activity clearly based in the “religious” civic sphere. We note, however, that a clear distinction between “religious” and “nonreligious” organizational volunteering is nearly impossible. First, congregations see organizing the laity to meet human needs in the community as central to their religious mission. Second, most churches have deaconate efforts designed to help congregants in financial need or in need of a job. This volunteering is organized “at and through” a congregation, but should we consider it as benefiting the religious community or as a contribution to the broader health of the local community? In these and other cases, we would need a keen philosopher to divide the religious from the community benefit.

A further problem is that the congregational volunteering question mentions “helping others through programs organized by your place of worship.” Some respondents may have interpreted that phrase to include volunteering through a church committee that serves at a local soup kitchen or visiting elderly shut-ins as a church leader. Others may not have been quite sure where to report their mission trip organized through the church but involving service to the poor in rural Appalachia. The wording of the questions indicate that volunteering for parachurch organizations such as parish schools, or religiously affiliated nursing home facilities would not qualify as church-based, but we cannot be certain that every respondent fully appreciates such a distinction.

Despite these measurement problems, the PSID volunteering module provides an adequate, if imperfect, measure of volunteering for the benefit of a religious community. While the other categories of volunteering include activity through non-congregational religious organizations, the wording focuses the respondent on volunteer activity designed to meet human needs. Thus we will refer to activity in these organizations as volunteering in the “nonreligious” civic sphere.

We used responses to these questions to test nearly 40 outcomes related to volunteering and philanthropy. First, we calculate two broad and six detailed volunteerism measures indicating (1) whether or not respondents volunteered for any organization or cause, and if so, (2) if they volunteered at or through a religious congregation, or (3) a non-congregational organization (defined as activity in any category other than a congregation). We also look at five specific types of non-congregational volunteering: youth, seniors/ elderly, healthcare, poverty, social change, and other community organizations. Our first measure is an indicator of broad engagement in civic life, while the subsequent measures allow us to test whether volunteerism is oriented in or through a faith community or clearly involves engagement in the community beyond religious congregations. We also consider (4) the total number of organizational types (or charitable causes) reported by the respondent as organizations through which they volunteered.

We take a similar approach to the measures of philanthropy. First, we created (5) an indicator of whether the subject donated money to any cause or charitable organization in 2012. Then, we used data on the monetary value of donations to a variety of types of organizations and causes to (6) calculate the amount of money donated to all charitable causes and organizations in 2012. We also calculated measures of whether and how much respondents’ households donated to any of 11 types of causes: religious, mixed mission, poverty, healthcare, education, youth, arts and culture, community organizations, environmental causes, international aid and peace, or “other” causes such as Veterans services. These measures allow us to test sector effects on several patterns simultaneously: whether one donates, where one donates, the total dollar value of gifts, and the dollar value of donations to any given category of social concern.

Regarding the dollar value to each type of organization or cause, we run models with and without a control for the total dollars given to charitable causes by the respondent. The additional control variable, total donations, helps us shed light on how important a particular cause is in the respondent’s giving “portfolio.” That is, net of how much the respondent donates in total, how much does the respondent donate to a particular cause? We should be clear that this approach measures the relative priority of a cause based on the dollar value of donations to the cause relative to the total amount of charitable donations. Note that this approach is not always the best measure of donation priorities. One might argue, for example, that giving 50 percent of one’s total donations to a congregation shows relative disregard for an Evangelical congregation if that percentage does not constitute at least 10 percent of income, which would satisfy the Evangelical obligation of tithing. In contrast, a small amount sent to the Sierra Club may indicate a high level of commitment to environmental causes even if it is not a high percentage of total donations, especially for a large donor. Or does it? Should we focus on the absolute or the relative value of donations to the Sierra Club? Depending on the exact question about donation priorities (how much do you give to a cause, or how much of your total donations do you give?), the findings with or without the control for total donations could be useful. In this paper, we consider both models.

We also created measures that will let us test for broader patterns in philanthropic giving. As with volunteerism, we are interested in whether congregations or other community groups and causes benefit from any observed sector differences in benefaction. For this, we calculated (7) the percentage of income tithed to a faith group or congregation, and (8) the value of financial support provided to non-congregational groups and causes.

PSID has a number of unparalleled strengths for our analysis. In survey research, respondents often provide socially desirable responses or opt out of questions that are sensitive in nature (Tourangeau and Yan 2007). PSID respondents are less likely to do so since they have developed a relationship with the study as part of a multi-decade longitudinal panel and are primed to discuss sensitive topics during in-person interviews. In other words, PSID is uniquely positioned to reliably collect sensitive information, such as income and philanthropy, because respondents are used to talking to a PSID interviewer about their finances every other year. Moreover, PSID takes the time and effort to collect precise figures on income and giving, and to accurately capture the extent of time and money donated to charitable causes and organizations.

The predictor variable of interest in our models is school sector. We compare differences in all 38 outcomes by the following high school sectors: public, Catholic, Evangelical Protestant (all non-Catholic religious school graduates in PSID), and non-religious private. A serious limitation of the PSID data available to us is the undifferentiated “non-Catholic religious” sector in the public PSID dataset. However, in the U.S. the proportion of non-Catholic religious school graduates that would not be Evangelical Protestant school graduates is very small, which should limit the impact of measurement error. Keeping in mind this limitation, we will let the non-Catholic religious category stand in for the Evangelical Protestant sector. Over 80% of adults in our sample who attended Evangelical Protestant schools identify as Evangelical Protestant as adults, which gives us some confidence that the vast majority of respondents in the non-Catholic religious sector are alumni/ae from Evangelical Protestant schools.

To test for sector effects, we run a series of regression models for each outcome, using school sector as a predictor of each generosity or volunteering outcome. This gives us the opportunity to see which behaviors are either more or less common for graduates of each sector in comparison with public school graduates. The specific method of estimation is determined by the distribution of the outcome of interest. For example, for indicator outcomes such as whether one volunteers, which only takes the value of 1 or 0, we use logistic regression, while ordinary least square (OLS) regression is used for continuous measures.

In our first model, we calculate the participation rate (for indicators like volunteering) or average (for measures of percent or amount donated) for adults who attended each sector. In subsequent models, we incorporate information into our estimates that accounts for the effect of demographic characteristics, education and income, or religion on the outcome of interest. This process allows us to either “isolate” the relationship between school sector and the outcome or “explain” why any sector differences we observe in our first model might not actually be a function of the types of schools that people attend but are due to other factors.

In some cases, we also discuss the possibility that school sector effects operate indirectly on the outcome of interest. Evangelical Protestant schools, for example, increase the likelihood that respondents identify as Evangelical Protestants in adulthood, which in turn increases the likelihood that individuals donate to congregations (Hoge et al. 1997; Smith, Emerson and Snell 2008a). Catholic schools likely increase the educational attainment of their graduates, and higher levels of education have strong effects on volunteering and giving (Musick and Wilson 2008).

Results

Volunteerism

The first model assesses the relationship between high school sector and volunteering at or through an organization. We find that graduates of Evangelical Protestant schools and other non-Catholic religious schools are about 40 percent more likely than public school alumni/ae to volunteer through at least one organization (Figure 1). Catholic school graduates are also more likely to volunteer than public school graduates, but this is accounted for by the higher average socioeconomic status of Catholic school graduates. The extent to which Catholic schools impact the volunteering of their graduates, then, depends on the impact of the school on increasing the educational or income level of their graduates. The nonreligious private school graduate is no more or less likely to report organizational volunteering than a public school graduate after including socioeconomic controls.

The next series of models consider the types of organizations through which PSID respondents volunteer. Since we are interested in the distribution of volunteering across religious and nonreligious civic organizations, we limit our sample to those who reported doing some volunteering through an organization. Among these volunteers, we would expect that religious school graduates would be more likely to volunteer through religious organizations. What we find is that Evangelical Protestant school graduates are more than twice as likely as public school graduates to volunteer at or through a religious congregation (Figure 2). This difference holds up even after accounting for current religious tradition and frequency of religious service attendance, which is consistent with the much higher levels of commitment to churches found among Evangelical Protestant school graduates (Pennings et al. 2011, Pennings et al. 2014). Interestingly, Catholic school graduates are less likely to volunteer through a church than are public school graduates, but this is not statistically significant after accounting for demographics and the higher socioeconomic status of the Catholic school graduates.

Our next model considers volunteering for organizations involved in youth causes, including sports and grade schools. Catholic school graduates appear to be much more likely to take part in youth volunteering, but this is accounted for by the (higher) socioeconomic status and Catholicism of these graduates. Interestingly, Catholic adherents compared to the nonreligious in the PSID are more than twice as likely to volunteer in youth organizations. To the extent that Catholic schools increase the likelihood that graduates become or remain Catholic, they have an important impact on increasing volunteering with youth-focused organizations among their students. Nonreligious private school graduates show a much higher level of youth volunteering than public school graduates, but these effects are not statistically significant in our sample. Although we would expect that Evangelical Protestant school graduates have a strong inclination to volunteer for youth causes, this must be channeled through congregations, as we find no statistically significant difference between Evangelical Protestant and public school graduates on volunteering at or through a youth organization.

Turning to volunteering for organizations involved with senior citizens or health care, we do not find significant school sector effects. This is somewhat surprising given that Catholic schools likely have established bridges to the large number of Catholic health care institutions, which could facilitate volunteering among their students and alumni/ae. The results do reveal a positive effect of Catholic school attendance on volunteering for a health care organization, but these findings are not statistically significant in this relatively small sample of around 2,000 volunteers.

We also consider whether school sector matters for the likelihood of volunteering for an organization dedicated to helping those in poverty. Interestingly, Catholic school graduates are over 50 percent more likely to volunteer for these organizations than are public school graduates, even after including demographic, SES, and religion controls (Figure 2). Evangelical Protestant school graduates are about 40 percent less likely to volunteer for this type of organization compared to public school graduates. If we include the full sample of volunteers and non-volunteers, we find the same negative effect for Evangelical Protestant and other non-religious school graduates, but this result is not statistically significant, and appears to be due to the current religious tradition and higher levels of religiosity of such graduates. Since Evangelical Protestant schools likely have some positive impact on the religious tradition and religiosity of their graduates (Pennings et al. 2011, Uecker 2009), after accounting for the influence these schools have on commitments later in life, they likely have a negative indirect effect on volunteering for organizations working to help those in poverty. Among volunteers, in contrast, the evidence is clear that Evangelical Protestant school graduates are less likely than public school graduates to be involved in poverty relief efforts outside of congregations.

Congregational v. Non-Congregational Volunteering

Next we grouped all volunteering in non-congregational organizations together to consider whether school sector is related to volunteering for any of the non-congregational organizational types offered in the PSID survey1 . We find that Evangelical Protestant school graduates are just over 50 percent less likely than public school graduates to volunteer in organizations outside of their congregations. This could reflect crowding out processes, in which time dedicated to volunteering at or through a congregation reduces the opportunity to donate time elsewhere (Wuthnow 1999). Alternatively, Evangelical Protestant and other non-Catholic religious school graduates do take up causes of youth, poverty, and health care, but these efforts are organized through their congregations. Either way, these findings indicate that volunteering among adults who attended Evangelical Protestant school tends to be organized at or through their religious congregation.

To flesh out the findings for congregational versus non-congregational volunteering, we include models that directly test whether high school experiences encourage Americans toward one of the following volunteering “portfolios:” 1) congregational volunteering only, 2) non-congregational volunteering only, or 3) both congregational and non-congregational volunteering. In separate (multinomial) models, we consider the choice of these portfolios both with and without a fourth group--(4) those who do not volunteer at or through an organization. We find that Evangelical Protestant school graduates, compared to public school graduates, are more than two and a half times more likely to volunteer at or through a congregation than not to volunteer at all. This holds even after accounting for religious tradition and religiosity. Nonreligious private school graduates are more likely to be found among the congregational volunteers, but this effect is not statistically significant.

The contrast between “no volunteering” and non-congregational volunteering reveals no strong school sector effects on pushing their graduates toward one specific portfolios of organizational volunteering. Compared to public school graduates, Catholic school graduates are about 40 percent more likely to be non-congregational volunteers rather than non-volunteers, but this effect is not statistically significant after accounting for socioeconomic status. Considering the contrast of volunteering both in a congregation and in a non-congregational organization, we find that graduates of Evangelical Protestant and other non-Catholic religious schools are about 65 percent more likely to be omnivorous volunteers than to stay at home. Although this effect is not as substantial as it is when the choice is to volunteer exclusively at/through a religious congregation, it is quite strong, indicating that, on average, Evangelical Protestant school graduates are not particularly averse to spreading their volunteering activities across religious and nonreligious organizations. Catholic school graduates and especially nonreligious private school graduates show this same preference for religious and nonreligious organizational volunteering, but the results are not statistically significant after accounting for demographics and other factors.

After limiting our sample to those who do some volunteering at or through organizations, our findings lead to very similar conclusions. Graduates of Evangelical Protestant and other non-Catholic religious schools are well over two and a half times more likely to choose to volunteer through a congregation compared to volunteering at a non-congregational organization. Catholic school graduates are about 50 percent less likely to choose congregational volunteering than are graduates of public schools, but this is accounted for by socioeconomic status. Compared to public school graduates, graduates of Evangelical Protestant schools are also about 70 percent more likely to volunteer in congregational and non-congregational organizations rather than exclusively in non-congregational organizations. Catholic and private nonreligious school graduates also favor a diverse volunteering portfolio, but these effects are not statistically significant. Changing the comparison group to congregational volunteers, we see that Evangelical Protestant school graduates are about 40 percent less likely to volunteer both in and outside of congregations. But this effect is not statistically significant after controlling for religiosity and religious affiliation. Among volunteers, then, graduates of Evangelical Protestant and other non-Catholic religious schools are not likely only to volunteer for non-congregational causes, but they are willing to cross the religious/ nonreligious boundary at the same rates as public school graduates.

Another approach to the distribution of volunteering within the civic sector is to consider how many volunteer activities are mentioned by the respondent. After adding up all the positive responses to each of the PSID survey questions about various types of organizations, we do not find strong relationships between school sector and the number of organizations that respondents volunteered for. However, after controlling for religious affiliation and attendance, which is likely important for Catholics since they tend to have fewer volunteering opportunities within the parish than do Protestants (Verba, Schlozman and Brady 1995), we find that Catholic school graduates are volunteering for a slightly greater number of organizations on average than are public school graduates. As we discuss below regarding donation “omnivores,” this finding may reflect the Catholic school emphasis on service to humanity, which includes a broad religious obligation to meet human needs and attempt to “bind the wounds,” to use biblical language, of victims of social injustice.

Financial Donations

The PSID’s strength is measurement of income and wealth across generations. In addition, a fairly comprehensive module on “philanthropy” provides one of the best U.S. assessments of where charitable dollars are directed. And PSID is one of the few surveys that is both large enough in sample size and includes information on high school experiences. We start with an assessment of whether and how much Americans from different high school sectors donate to charitable causes (not including political donations), and then turn to investigating how school sector affects charitable priorities.

We first consider whether Americans made any financial donation greater than twenty-five dollars. The results clearly show the religious school advantage in giving (Figure 3). Catholic school graduates and Evangelical Protestant school graduates are more likely than public school graduates to make charitable donations even after accounting for differences in socioeconomic status and religion (Figure 3). Private nonreligious school graduates are also more likely to make financial donations, but this is accounted for by demographic differences between private nonreligious and public school graduates. And after accounting for socioeconomic differences, the private nonreligious and public school graduates are nearly identical in the likelihood of making a charitable contribution.

The estimates of total charitable giving are even more striking. Graduates of Evangelical Protestant and other non-Catholic religious schools give on average $1,273 more than public school graduates to charitable causes, net of the other variables in the model (Figure 3). For comparison, the difference between those who attended an Evangelical Protestant school and those from public schools is about $150 greater than the average difference between someone with a bachelor’s degree relative to someone who does not have any postsecondary education. What is particularly surprising about the size of this difference is that it appears to be entirely the result of attending an Evangelical Protestant (or other non-Catholic religious) high school, rather than simply the result of current religious identification and behavior. To the extent that Evangelical Protestant schools encourage students to stay in the faith or attend church more regularly (Pennings et al. 2011), the estimates of the financial donations of graduates of Evangelical Protestant schools would be even higher. We also consider a slightly different approach to financial donations by constructing a variable for percent of income dedicated to charitable organizations and causes. When predicting percent of income donated to charity, we find that Evangelical Protestant school graduates are on average 4.16 points higher on this scale than public school graduates, even after controlling for religious attendance and affiliation. In terms of the proportion of income dedicated to charitable causes, graduates of non-Catholic religious schools, who give 4 percent of income more than graduates of other school sectors, is an extreme outlier. No other school sector is significantly different on this outcome.

Where does that money go? How do giving priorities vary across school sector? The PSID provides a reasonably detailed list of charitable organizations, and asks respondents whether and how much they gave to each. These include variables for religious, youth and family, poverty relief, education, and health care. We discuss results from three models for each of these variables: 1) whether respondents give to a particular cause or not, 2) the total dollar amount dedicated to each cause, and 3) the priority of a cause in their “portfolio” of charitable donations. The latter is also measured by the dollar value of donations to a cause, but the models include a control for total charitable donations.

We consider first donations for “religious purposes or spiritual development,” including religious congregations. Graduates of both Catholic and non-Catholic religious schools are much more likely to give to a religious cause than are graduates of public school graduates (Figure 4). However, the Catholic school effect is explained by socioeconomic status (SES) differences; we have no clear evidence of the impact of Catholic school on religious giving after including controls for SES. Even after accounting for the current religion of the respondents, Evangelical Protestant and other non-Catholic religious school graduates are nearly 1.8 times more likely to make a religious donation than are public school graduates. Private nonreligious school graduates are no more or less likely to donate to a religious or spiritual cause than are public school graduates.

When considering the total dollar amount of donations to religious causes, we find that, among donors, Catholic school graduates donate fewer dollars to religious causes—a finding that holds until we account for the lower levels of religious giving among self-identified Catholics. Net of religious identity, Catholic school graduates give a slightly higher amount to religious causes. But this difference is not statistically significant. When considering the relative priority of giving money to religious causes versus other causes, we find—after controlling for total donations—that Catholic school graduates donate about $250 dollars less than public school graduates to religious causes. But again this effect nearly disappears and is not statistically significant after accounting for current religion and religiosity of the respondent. It appears that the difference is explained by the relatively low levels of giving among American Catholics generally.

In contrast, graduates of Evangelical Protestant and other non-Catholic religious schools give nearly $1,100 more to religious causes—and that is above and beyond the positive effect on religious giving of Evangelical Protestants and those who attend church more (Figure 4). To shed further light on this finding, we considered a slightly different dependent variable, the percentage of income donated to religious and spiritual development causes. Tithing, that is, donating 10% of one’s income to church, is uncommon, according to the evidence available in the PSID. But 33% of PSID respondents donated at least some money in the previous calendar year to a religious organization. Regarding the percentage of income given to religious and spiritual development, after limiting our sample to donors, we find that only graduates of non-Catholic religious schools are significantly different from public school graduates, donating a higher percentage of earnings to religious organizations than any other sector– about 2.2% higher on average. Given that a relatively generous donor in the US gives 3-5 percent of income to charity (Smith, Emerson and Snell 2008b; Smith and Davidson 2014), the Evangelical Protestant school effect on congregational giving is remarkable.

Moving to the relative priority of religious donations, we find that, net of income and total donations, graduates of Evangelical Protestant and other non-Catholic religious schools are giving about $350 more to religious causes than are public school graduates, but this is largely accounted for by socioeconomic and religious differences. That is, we do not find a statistically significant effect of non-Catholic religious schooling on the relative priority of religious causes compared to other causes. The effect of non-Catholic religious school attendance on the priority of religious donations would be indirect, through their effect of religious identification and on higher religiosity.

When considering differences among Americans who make at least some charitable donations, we do not find a statistically significant effect for Evangelical school graduates on increasing the place of religious donations in a giving portfolio.

Next, we examined sector differences in donating to organizations that addresses diverse charitable causes, such as United Way, Catholic Relief Services, and a local community organization. After controls for current religion, Catholic school graduates are about 34 percent more likely to give to this type of organization than are public school graduates. That estimate jumps to 56 percent if we don’t account for religious differences. Evangelical Protestant school graduates are no more or less likely than public school graduates to make a donation to organizations like the United Way. Graduates of private nonreligious schools, like Catholic school graduates, are much more likely than public school graduates to give to a charitable organization that serves multiple causes.

When limiting our sample to donors, and estimating the total dollar value of contributions to general purpose charitable organizations, we find that Catholic school graduates give more than graduates of public schools to these organizations, but this difference is not statistically significant after accounting for socioeconomic status. Private nonreligious school graduates give more, and Evangelical Protestant school graduates give slightly less, both in comparison to public school graduates, but again these differences are not statistically significant in this sample. In our estimate of the proportion of total charitable donations that flow to general purpose organizations, we find that Catholic school graduates are directing relatively more dollars to these organizations, but this tendency is accounted for entirely by socioeconomic and religious differences. After accounting for total donations and the rest of our socioeconomic and religious variables, we find that graduates of Evangelical Protestant and other non-Catholic religious schools are significantly less likely to give to general purpose charitable organizations than are public school graduates—about $125 less on average. Private nonreligious school graduates appear to prioritize these organizations slightly more than public school graduates, but this effect is not statistically significant in our models.

We find significant sector effects in the likelihood of donating to poverty relief. Specifically, graduates of Catholic schools are forty percent more likely to give to poverty relief programs compared to public school graduates (Figure 4). Graduates of non-Catholic religious schools have the same likelihood of giving for poverty relief as do public school graduates. Private nonreligious school graduates are more likely to give to organizations working for poverty relief, but this difference is not statistically significant after accounting for socioeconomic differences. In dollar value, religious school graduates give slightly more to poverty relief than public school graduates, but this difference is accounted for by socioeconomic status (for Catholic school graduates) and religious identification and behavior (for non-Catholic religious school graduates). The positive effect of Evangelical Protestant schooling on poverty relief dollars would have to be indirect through a schooling effect on maintaining an Evangelical Protestant identity. In terms of the relative amount directed to poverty relief, we find again that Catholic school graduates are more likely to put a higher proportion of their donation dollars to poverty relief, but this effect is not statistically significant after accounting for socioeconomic differences.

In addition, health care or medical research donations are popular with Catholic school graduates, who are almost 30% more likely to give to health care causes than are public school graduates, even after including religion controls (Figure 4), although we find no significant sector differences in terms of dollar value donated for this type of research. When predicting the relative importance of health care giving in respondents’ donation “portfolios,” religious school graduates are not particularly focused on health care donations compared to graduates of public school. Evangelical Protestant and other non-Catholic religious school graduates direct significantly fewer dollars to health care or medical research relative to their total donations than do public school graduates (about $60 less after controlling for total donations and other variables in the model).

The findings for the relation of school sector and giving to educational causes are nearly identical with those for poverty relief. Graduates of Catholic schools are over 50 percent more likely to donate to educational causes than are graduates of public schools (Figure 4). Non-religious private school graduates have a greater propensity to donate to educational causes, but this is accounted for by socioeconomic differences. Evangelical Protestant school graduates are no more or less likely than public school graduates to give to educational institutions. The total dollar value of educational donations is higher among Catholic school graduates, but this is accounted for by socioeconomic differences. After including all controls, we find the non-Catholic religious school graduates give almost $250 more to educational causes than public school graduates, on average, and net of the other variables in the model. When we consider giving portfolios, the proportion directed to educational causes does not vary by school sector. Catholic school graduates do give relatively more than public school graduates to educational causes, but this difference is not statistically significant after accounting for socioeconomic status.

Donations to youth and family services are also attractive to Catholic school graduates, who are about 46 percent more likely to donate to these causes (Figure 4). They give higher dollar amounts on average (about $24 more) to youth and family services than public school graduates as well. In terms of giving priorities, however, we do not find evidence of school sector effects. No school sector gives more or less to youth and family causes relative to their total financial contributions.

The findings regarding giving to arts and cultural organizations are particularly surprising. We expected graduates of Catholic schools to be significantly more likely to make this kind of donation than public school graduates, which is precisely what we find. The Evangelical Protestant school graduate findings are unexpected, though. Other research has shown lower levels of involvement of Evangelical Protestant school graduates in arts and cultural organizations (Pennings et al. 2011). But the philanthropy data in the PSID reveals that Evangelical Protestant and other non-Catholic religious school graduates, after accounting for socioeconomic and religious differences, are nearly twice as likely as public school graduates to make donations to these causes. Perhaps not as surprising is the interest in arts and cultural organizations among the private nonreligious school graduates. Alumni/ae of private nonreligious schools are over two times more likely than public school graduates to make a donation to an arts or cultural organization, on average, and net of the other variables in the full model. Regarding the dollar value of contributions to arts and cultural organizations, in contrast, we do not find any school sector differences, either in terms of dollar amounts or the relative priority of arts and cultural causes in Americans’ giving portfolio.

The PSID philanthropy module offered to respondents a catch-all category for donations to other causes or organizations, and followed up to get some sense of the type of charitable cause. We grouped several of these causes under the umbrella of “community organizations,” which include organizations to improve neighborhoods. Catholic school graduates are particularly likely to donate to these groups, but this difference between Catholic and public school graduates is not statistically significant in our full models. The PSID data do not reveal any school sector differences in the dollar value of community organization donations, nor do we find any school sector differences in the relative importance of these groups in one’s giving portfolio. We speculate that this may reflect the relative size and scope of these organizations, for which even relatively small donations have a large impact.

Although we don’t consider donations for political causes—the PSID module explicitly asks respondents to ignore these donations—most environmental organizations are very close to the line that supposedly divides charitable causes and political issues and action. Perhaps for this reason, our findings regarding donations to environmental causes are very crisp. Catholic and private nonreligious school graduates are much more likely to contribute to an environmental cause than are public school graduates. The likelihood is nearly 80 percent higher for private nonreligious school graduates. The likelihood that Evangelical Protestant school graduates give to these causes is lower than the likelihood for graduates of public schools, but this difference is not statistically significant.

The amounts given are quite similar across school sectors, except for the strong positive effect for nonreligious private school graduates giving to environmental causes. On average, even after controls, graduates of nonreligious private schools give nearly 40 dollars more than those from public schools to these causes (Figure 4). The relative importance of these donations in a respondent’s giving portfolio is equally fascinating. Catholic school graduates channel relatively more financial donations to environmental causes, but this is not significantly different from public school graduates after controls, perhaps because, as shown above, Catholic school graduates have a multitude of charitable causes that concern them. In contrast, nonreligious private school graduates place a very high priority on environmental causes, at least as measured by the place of environmental causes in their giving portfolio, relative to public school graduates, on average and net of the other variables in the model, including socioeconomic status and religion.

In terms of the relative priority of environmental giving compared to other types of charitable donations, Evangelical Protestant school graduates direct significantly fewer dollars toward environmental causes than public school graduates, at least as compared to the share of their total charitable dollars that go toward environmental causes (Figure 4). But this difference is only significant before controlling for religion. If Evangelical Protestant schools influence their students to give less to environmental causes when they are adults, the effect would have to be indirect, through a positive influence on remaining in Evangelical Protestant churches and attending church regularly. We should be clear that this analysis measures relative priority based solely on what is essentially the proportion of total financial giving that is directed toward environmental causes. We can only say that in terms of the relative importance of environmental causes in the Evangelical Protestant school graduates’ giving portfolio—measured in dollars contributed, which is very high—we find a lower commitment to giving to environmental causes.

A similar pattern is evident for international aid and world peace donations. Catholic and nonreligious private school graduates are much more likely to donate to these causes than are public school graduates, even after controls (Figure 4). Non-Catholic religious school graduates are less likely to give in this way, but this effect is not statistically significant. In terms of the total dollar amounts or the place of international giving in respondents’ portfolios, we do not find any school sector differences.

Donations to Religious v. Non-Religious Organizations

Next we group all the nonreligious charitable causes into one, and consider donation portfolios that include 1) only religious, 2) only nonreligious, and 3) both religious and nonreligious donations. This analysis will shed light on the role of the religious/nonreligious boundary in shaping giving priorities. Consistent with our findings when each cause is considered separately, we find a strong likelihood that Catholic school graduates give to nonreligious causes, rather than not giving at all or giving only to religious causes (Figure 5). Private nonreligious school graduates are also more likely to give to nonreligious causes, but this is not statistically significant after accounting for socioeconomic differences. Evangelical Protestant and other non-Catholic religious school graduates are not different from public school graduates on this outcome.

When predicting the number of dollars directed to nonreligious causes, we limit the sample to those who make charitable contributions. In terms of total dollar value, graduates of all the private school sectors have higher donation amounts than graduates of public schools, Catholic school graduates in particular, but these differences are not statistically significant in this sample (Figure 5). We also consider the number of dollars relative to the total charitable donations made by the respondent. Among donors, we find that Catholic school attendance is strongly and positively related to giving a relatively higher amount to nonreligious causes, but this is accounted for by the religion variables. In other words, the Catholic school and public school donors are no different in the size of their contributions to nonreligious causes after we account for religious differences. The Catholic school effect on nonreligious donations may operate indirectly through the school effect on adult religious outcomes. Evangelical Protestant school graduates donate relatively fewer dollars to nonreligious causes, which we can attribute to socioeconomic status. Private nonreligious school graduates are no different than public school graduates in the number of dollars directed to nonreligious causes.

Another approach to understanding donation priorities is to ask which respondents donate to a variety of causes rather than one or two. Giving to multiple causes may reflect the extent to which donors are interested in various charitable causes, rather than focused on one, and broadly networked to diverse charitable organizations and opportunities. The cultural “omnivore” may have a corollary in civic life (Hustinx et al. 2012, Peterson and Kern 1996, Lizardo 2006, Lizardo and Skiles 2008). We can think of the donor “omnivore” as both broadly attuned to human and social needs and centrally networked in the civic sphere. When predicting the total number of causes offered in the PSID module that received donations from a respondent, we find that nonreligious private school graduates give to a significantly higher number of causes than public school graduates, which may reflect the elite social position of graduates of nonreligious private schools, and their willingness to consider both religious and nonreligious charitable giving. Catholic school attendance is also strongly and positively related to giving to more causes. Most likely this reflects the openness of Catholics to religious and nonreligious charitable organizations as well as a broad concern with human needs and social injustice both in their local communities and overseas. Perhaps another way of interpreting these findings is that graduates of Catholic schools are not likely to meet a charitable solicitation they don’t like— or feel an obligation to respond to with at least a small donation. Overall, our findings on donation priorities seem consistent with the Catholic sense of obligation to humanity, which likely is taught, incorporated into student volunteering opportunities, and modelled by teachers within Catholic high schools.

Finally, we expect that the impact of religious schools on giving and volunteering varies depending on whether the respondent was very committed to their faith when in high school and remain committed into their adult years. The PSID does not have detailed measures of religion, but it does provide the current frequency of religious service attendance for each respondent. The question is whether the giving patterns of, say, Evangelical Protestant school graduates are different if we take into account the fact that some graduates of Evangelical Protestant schools are highly committed to church attendance while others are not. Perhaps the Evangelical Protestant school effect is “activated” more strongly among alumni/ae who have high levels of religiosity. That theoretical argument is reasonable, but we find no evidence for it in the PSID data. In our models, neither Catholic nor Evangelical Protestant school effects depend on the current level of religious attendance. We can only conclude that the effect of religious school attendance does not depend on whether the respondent is highly religious or not; it applies equally to students from all levels of religiosity. This conclusion is not definitive, however, since we are not able in these analyses to consider a richer set of religion variables at different points in the life course.

Summary and Conclusion

Many prominent democratic theorists are skeptical that private education, especially in religious schools and homeschooling, can prepare students for active engagement in civic life (Reich 2002). One of the challenges for research on religious schools and civic engagement is the quality of data available to link religious school experiences and adult behaviors and orientations. Several adult cross-sectional research designs have helped to meet these challenges (Pennings et al 2012, 2014; Campbell 2001), but suitable longitudinal surveys are scarce since most do not have information on high school experiences or the appropriate measures and a large enough sample size to estimate effects of smaller religious school sectors. The Panel Study of Income Dynamics is one of the largest and most respected longitudinal surveys available, but has been neglected in research on religious schools and civic preparation. Our analysis of the PSID reveals a strong connection between religious schools and civic involvement, which should allay concerns that religious schooling is harmful to a flourishing civil society.

We find a generally positive influence of Catholic schooling on civic engagement. Despite mixed results in other research (Dee 2003a, Dill 2009, Hill and den Dulk 2013, Pennings et al. 2011, Pennings et al. 2014, Sikkink 2012), the PSID analysis reveals a much higher propensity among Catholic school graduates to make charitable donations. Catholic school alumni/ae are active volunteers as well, though the extent that this is a function of their Catholic school experiences depends on accounting for the impact of Catholic schools in improving the educational attainment of their graduates, which in turn leads to a greater likelihood of volunteering.

What is particularly striking is the pattern of volunteering and giving among Catholic school graduates, which generally reflects the approach to Catholic social teaching that is a significant part of the organizational culture of most Catholic high schools. The importance of giving to poverty relief, health care, education, youth and family services, and multi-purpose charitable organizations, such as United Way and Catholic Relief Services, reflects traditional Catholic social concerns as well as organizational ties between Catholic schools and civic organizations. The higher dollar amount given to youth and family services by Catholic school graduates furthers the claim that Catholic schools form students with distinctive Catholic civic concerns. Moreover, the place of international aid donations among Catholic school graduates likely reflects the global emphases and networks of Catholicism, which are built and fostered within Catholic schools. The likelihood of Catholic school graduate donations to arts and cultural as well as environmental organizations, which are closely intertwined with class divides in the US, is perhaps the only evidence in our findings for the “eliting” of the Catholic schools (Baker and Riordan 1998). But these concerns have a religious motive within Catholicism as well. Focused attention in “progressive” Catholic schools on the arts and the environment, along with poverty and other issues of social injustice, constitute an important part of Catholic school mission and identity. Catholic school students seem to be heeding the message and taking up their religious calling in the civic sphere. The strength of the Catholic school community and organizational ties to civic life likely strengthens the civic formation effects of Catholic schools.

This Catholic school civic education is not oriented solely toward religious organizations, and certainly not toward congregations. We find some evidence that Catholic school graduates are more attuned to volunteering and giving outside of the congregation; they certainly are willing and able to cross the boundary between the religious and nonreligious civic sphere. Catholic school graduates see and respond to diverse charitable causes, whether through religious or nonreligious organizations. Though on many of the PSID outcomes Catholic school graduates are not different from public school graduates after controlling for socioeconomic status, the overall portrait of Catholic schools reveals civic strengths and a small but significant Catholic school advantage in civic education.

Our findings reveal that Evangelical Protestant schools are forming citizens in significant and valuable ways, though the question remains whether the focus on the religious civic sphere is sufficient for a flourishing civil society. Evangelical Protestant school graduates are more likely to volunteer than public school graduates, even after we account for the higher levels of religiosity among Evangelical Protestant school graduates. And the impact of Evangelical Protestant schooling on the total amount of charitable donations is remarkable. The experience of a close-knit school community, combined with a religious culture that values putting others—especially the less powerful—before oneself and making sacrifices for the common good of the community, carries over into a strong commitment to volunteering and giving in civic life as an adult. Given the constraints of the PSID data, we do not have a complete understanding of how Evangelical Protestant school graduates giving and volunteering influences the broader civic sphere. Taking all our findings of higher levels of giving and volunteering among Evangelical Protestant school graduates, however, we conclude that the evidence is consistent with the claim that Evangelical Protestant schooling generates high levels of civic engagement of their graduates.

Our data show that Evangelical Protestant school graduates are highly committed to volunteering and giving in religious organizations, especially congregations. We cannot tell the extent that those donations of time and money contribute solely to the religious life of the congregation, such as volunteering to serve as an usher or choir member at the Sunday worship service. But we should not ignore the fact that religious organizations make up nearly 50 percent of the civic sphere in the U.S and Evangelical Protestant school graduates are very active in them. Congregations are vital organizational partners in broader civic activities and programs (Ammerman and Farnsley 1997, Becker and Dhingra 2001, Wuthnow and Hodgkinson 1990, Wuthnow 1999, Wuthnow 2009), and offer “latent” social capital that is drawn upon in times of crisis (Ammerman 2005). Nearly all mainline Protestant congregations, and an increasing number of Evangelical Protestant congregations, bring together religious and social welfare in their outreach programs, including in mission trips and local poverty relief efforts (Cnaan 2006, Wuthnow 2009). Moreover, some evidence suggests that Evangelical Protestant school graduates are not particularly averse to civic engagement outside religious organizations. Evangelical Protestant school graduates are more likely than public school graduates to volunteer both for congregational and non-congregational causes. And Evangelical Protestant school graduates compared to public school graduates are not more or less likely to volunteer outside of congregations rather than not volunteering at all. Evangelical Protestant school graduates are “joiners” in the sense of preferring to help rather than focusing exclusively on the self or family. We find no difference in the likelihood that Evangelical Protestant or public school graduates donate only to religious causes, rather than giving both to religious and nonreligious causes. Evangelical Protestant school graduates’ patterns of giving are not different from public school graduates in terms of the relative priority of religious versus nonreligious causes.

Besides the Evangelical Protestant school graduate focus on religious organizations, we do not find that Evangelical Protestant school graduates follow what is often thought of as the Evangelical script. They are not more or less likely to focus on youth or family services or education. There is some evidence that they do not particularly favor health care organizations compared to public school graduates. But more surprising is the higher likelihood that Evangelical Protestant school graduates will participate in arts and cultural organizations than public school graduates. While the evidence for Evangelical Protestant school graduates’ commitment to religious organizations is strong, the willingness to get involved in arts and cultural civic organizations pushes Evangelical Protestant school graduates beyond the religious civic sphere. Overall, we find that Evangelical Protestant school graduates have a strong commitment to giving and volunteering in religious organizations. Their commitment to civic engagement outside the religious civic sphere is not as high as Catholic school graduates, but in some important ways is similar to public school graduates.

Nonreligious private school graduates do not consistently stand out in our analysis of the PSID data. In general, they appear open to involvement in religious and nonreligious organizations. They do favor donations to multi-purpose organizations, such as the United Way. Consistent with expectations (Cookson and Persell 1985), nonreligious private school graduates strongly favor donations to arts and cultural organizations. The other strong distinction of nonreligious private school graduates is in volunteering and giving to environmental organizations. This may reflect the priorities of nonreligious schools, which have a history of promoting obligations to civic life but now find that “secular sacreds,” such as the arts and environmental causes, fit more easily into the culture of these schools and with the expectations of parents.

Although our findings provide fairly consistent evidence of a religious school advantage in giving and volunteering into adulthood, we do not conclude that public schools have entirely neglected their civic formation roots—at least in comparison to private schools in the U.S. If we limit our evidence to direct effects, public school graduates are as likely to volunteer for charitable organizations as nonreligious private and Catholic school graduates. Public school graduates are equivalent to the private sectors in volunteering for senior citizens and in giving to community organizations. They are statistically equal with private nonreligious school graduates in giving to multi-purpose organizations, such as United Way, youth and family services, and health care and medical research. There is no difference between public, Catholic, and nonreligious private school graduates in the propensity to volunteer outside of congregations versus not volunteering at all. Public school graduates’ likelihood of making a charitable donation is no less than nonreligious private school graduates after controls. And they do not neglect donations to religious causes relative to Catholic school graduates.

Compared to private schools, public schools face the vice grip of political accountability, special interest groups, management-union struggles, legal sphere impingement, and centralized governance (Arum 2003, Chubb and Moe 1990, Ravitch 2013, Ravitch 2016). In that context, school authority, alignment between peer and school culture, and a coherent civic mission are much more difficult to maintain (Damon 1996, Damon 2001, Grant 1988, Grant 2009). Under the social conditions we currently face, including a weak link between school and community (Coleman and Hoffer 1987, Grant 2009), we argue that religious schools have significant strengths in civic formation, but that in most respects public schools are not that far behind.

The limitations of this study point to necessary next steps. First, we do not yet have a good measure of Evangelical Protestant school attendance, and instead rely on a more general category of non-Catholic religious schools. Given the high proportion of Evangelical Protestant schools in this general category, the main findings should not be strongly affected by the lack of a direct measure of Evangelical Protestant schooling. Second, our models should do more to account for cross-generational strengths of the PSID, which would allow a direct assessment of family SES and religion effects rather than relying on the respondent’s current income, education, and religion. Third, the PSID has considerable strengths in addressing our questions, but does not have all the detail we would like to have on who was served by respondents’ volunteering and giving. While sector differences in philanthropy often result in resources flowing to churches, it is important to recognize that churches are not the only direct beneficiaries, since volunteering through congregations includes outreach within the local community or mission trips overseas that involve the provision of social services to those in poverty. Our findings, particularly on non-congregational giving, indicate that we need data collections that will allow closer investigation of sector differences in giving to non-congregational groups and organizations, which will help to sort out the impact of private religious schools on graduates’ nonreligious forms of philanthropy.

Fourth, an assessment of change over time in the relationship between school sector and civic engagement is warranted. While generally supportive, not all of the PSID findings in this paper are consistent with more recent cross-sectional surveys on school sector effects, such as the lack of a Catholic school on charitable donations and of an Evangelical Protestant schooling effect on involvement in arts and cultural organizations (Pennings et al. 2012). Discrepancies could reflect differences in methodology, since the PSID is a large national panel design, which does not depend on retrospective reports of high school sector. And differences in findings may be chalked up to measurement strategies, particularly the questions used to capture various forms of civic engagement. But a more likely explanation is the difference in the age of the respondents, since the PSID includes a sample of Americans of all ages while the Cardus Education Survey and the National Survey of Youth and Religion, for example, are limited to respondents in their young adult years. Additional studies that pay attention to cohort differences and to closely matching the modelling strategies across datasets will shed light on possible generational differences in school sector effects on adult outcomes. The religious school sectors in particular have faced many challenges since the 1970s that have likely changed the organizational culture of most religious schools (Baker and Riordan 1999, Baker and Riordan 1998, Dill 2012, Wagner 1997, Youniss and McLellan 1999, Youniss and Convey 2000). Further research is needed to chart these potential changes in the lasting influence of religious schools on the civic engagement of their graduates.